Information can be transferred between users at different locations using different types of communication networks. The information being transferred typically includes addressing information that indicates the location to which the information is to be relayed. This addressing information allows the network to route the information to the intended recipient.
The communications networks which transfer information include a number of interconnected nodes. A node is a location in the network at which incoming information, such as a stream of data packets, can flow out over one of a plurality of paths. A node contains sufficient intelligence to decide how to choose one of the paths based on the addressing information and the technology used in the network. Thus, the node decides how to route the information. The information may be routed to a next node, traversing the communication network, to arrive at the destination.
One such communication network is a conventional public switched telephone network ("PSTN"). In the conventional PSTN the node is a conventional telephone switch. The routing decision made by the node is for the setup of a telephone call. The addressing information on which the decision is based includes a callee's telephone number. Typically, the callee's telephone number is ten digits. These ten digits include a three digit area code and, in the United States, a three digit exchange. In a conventional PSTN, there is a strong correlation between phone numbers and geographic location. For example, in the ten-digit telephone number, the area code specifies a geographic region. The exchange specifies switches for a portion of the geographic region. The destination, the callee's telephone, is in this portion of the geographic region. In general, when more digits of a phone number are specified, the geographic location of the telephone switch which rings the callee's telephone is better determined. Thus, the nodes in the To conventional PSTN use the callee's telephone number to route the call directly to a geographic region close to the known physical destination of the phone call. Conventional telecommunications switches ensure that a call does not take an excessively long and inefficient path to the callee.
A communication network which utilizes computers is the Internet. Data packets can be transferred between two locations, a source and a destination, via the Internet. The Internet can be viewed as a collection of systems which are compatible with the Internet Protocol ("IP"). The Internet allows for a wider variety of services than other conventional communications networks such as a PSTN. The Internet should also be more fault-tolerant, and allow for a broad range of speeds and transmission media than a PSTN.
Each data packet transferred via the Internet includes addressing information. The addressing information in the data packet includes a destination address for the data packet's destination. The destination address includes an IP address of the destination. An IP address contains four integers, with the integers being separated by periods. Each integer is from zero to two hundred and fifty-five. For example, an IP address and, therefore, a destination address is 202.123.123.6.
IP addresses are typically assigned to autonomous systems ("ASs") within the Internet. An autonomous system is a collection of IP addresses and IP-capable devices and communications links under a single administrative domain. An autonomous system assigned the IP addresses may also be an Internet Service Provider ("ISP"). A particular AS may be assigned a range of IP addresses. For example, a single AS may have the IP addresses 202.123.4.0 through 202.123.7.255. In order to ensure that information is routed to the correct destination, each AS must be able to determine how to reach the autonomous system that is assigned the destination address. Therefore, the ASs of the Internet tell each other how to find every IP address range in the Internet. Each AS announces its range to the entire Internet, declaring that "if you have traffic bound for one of these IP addresses, send it to me." Other ASs use this information to ensure that the data packet reaches the AS assigned the destination address and, therefore, the destination.
There are multiple routes which a data packet can use to traverse the Internet. A route includes the ASs that a data packet travels through in order to reach the destination. Typically, a router acts as a node through which the data packet can access multiple routes to the destination. Using the IP packet's destination address, the routers within the Internet determine how to route the data packet to the AS assigned the destination address and, therefore, the destination. A data packet may pass through a series of routers before reaching the destination. Certain conventional routers select the route such that the fewest number of ASs are crossed before reaching the destination.
Although the Internet is capable of transmitting data packets from a source to a destination, the routing may be slow or may result in data packets being lost. There is no geographic correlation between the destination address and the physical location(s) associated with the AS assigned the destination address. IP addresses may come in clusters because a range of IP addresses may be assigned to the same AS. For example, an AS in Japan might be assigned the IP address range of 202.123.4.0 through 202.123.7.255. However, no convention specifies that all IP addresses of a particular range will be in an AS associated with a specific country or other geographic region. Therefore, similar destination addresses could correspond to destinations which are widely separated geographically. Nothing about a destination address implies the geographic location of the destination or the geographic location of the AS that is assigned the IP address of the destination.
In part because destination addresses are not linked to geographic areas, routers can make non-optimal routing decisions. Non-optimal routing may slow down the end-to-end Internet traffic stream or route the traffic over congested links. As discussed above, routers use the destination address in order to determine routing of a data packet. However, the destination address contains no information relating to the geographic area of the destination, the geographic area of the AS for the destination, or the geographic area of the routers between the source and the destination. As a result, a router may select a route that forces a data packet to travel a very large distance geographically. For example, to reach a destination having an IP address assigned to a first AS in Great Britain, a data packet originating in a second AS in the United States may travel to a third AS in Asia, then to the first AS in Great Britain because this route has the fewest ASs. Because the data packet travels such a large distance, transmission of the packet may be slowed. In addition, the data packet may cross a number of international links, which typically have a very high link utilization. As a result, the data packet may be lost or the transmission of the data packet may be slowed.
Accordingly, what is needed is a system and method for more efficiently routing packets. The present invention addresses such a need.